We have developed several novel approaches for analysis of massively parallel gene expression datasets. First, StepMiner, a tool that identifies step-wise transitions in the time course microarray datasets. Second, BooleanNet, a method of discovering Boolean implications between genes using these large numbers of gene expression datasets. Recently, we published a new method called MiDReG (Mining Developmentally Regulated Genes) that uses Boolean implications to successfully predict genes in developmental pathways. By initially applying this approach to lymphocyte differentiation, we discovered previously unrecognized markers for B cell differentiation, as well as a novel branchpoint of B cell and T cell development. The proposed project will build on our successful prediction of human B cell developmental genes using MiDReG to develop a general method for discovering cancer stem and progenitor cells. I am planning to validate this approach in human bladder cancer (Transitional Cell Carcinoma), first, because it is a simple model cancer to test, and, second, because our laboratory has the expertise to isolate and test cell populations for tumor-initiating potential. This method will be optimized for the discovery of stem and progenitor cells in bladder cancer and hopefully it will serve as a starting point for similar studies in other types of cancers. More than 90% of human bladder cancers arise from a simple epithelial tissue called urothelium, and are commonly called transitional cell carcinomas (TCC). Furthermore, human bladder cancer is fifth most common cancer in the United States. Our laboratory has established a working model for the xenotransplantation of human bladder cancer in mice and has recently discovered a tumor-initiating population in bladder cancer. Recent studies show that cancer is heterogeneous and forms a hierarchy of original tumor cell populations. However, a detailed bladder cancer developmental hierarchy remains unknown. My primary near-term goal in this proposal is to understand this hierarchy of bladder cancer cells using a systems biology approach to predict (diagnostic/prognostic) genes that mark specific populations. I will validate this approach using human cancer tissue microarrays in collaboration with Dr. Matt van de Rijn, and using xenotransplantation in collaboration with Drs. Robert Chin and Jens-Peter Volkmer. In the longer term, I will extend the method to identify stem and progenitor cells in other types of cancers during my independent investigator phase.